A lead researcher explaining the benefits of Cognition.run

A Comprehensive Guide to Behavioral Science

What behavioral science covers

Behavioral science is the umbrella term for disciplines that study what people (and other animals) do, using systematic observation and experiment. It includes experimental and cognitive psychology, behavioral economics, parts of neuroscience, sociology and anthropology, and applied fields such as human factors and public health. What unites them is a commitment to measuring behavior rather than relying on introspection or theory alone.

A short history

Modern behavioral science is usually dated to John B. Watson's 1913 paper "Psychology as the Behaviorist Views It", which argued that psychology should abandon the study of consciousness and restrict itself to observable behavior. Watson drew on Ivan Pavlov's work on conditioned reflexes in dogs, published in the preceding decade, which showed that a neutral stimulus paired with food comes to elicit salivation on its own.

B. F. Skinner extended this program from the 1930s onward with operant conditioning: behavior is shaped by its consequences, and schedules of reinforcement produce characteristic patterns of responding. Skinner's methods, including cumulative records and single-subject designs, remain in use in applied behavior analysis.

The so-called cognitive revolution of the 1950s and 1960s brought mental processes back as legitimate objects of study, provided they were inferred from behavior under controlled conditions. George Miller's 1956 paper on the capacity of short-term memory ("The Magical Number Seven, Plus or Minus Two") and Ulric Neisser's 1967 book Cognitive Psychology, which gave the field its name, are the usual landmarks. Reaction-time methods, signal detection theory and information-processing models date from this period.

From the 1970s, Daniel Kahneman and Amos Tversky documented systematic departures from rational choice: heuristics and biases in judgment (1974) and prospect theory (1979), which describes how people evaluate gains and losses relative to a reference point. Richard Thaler built on this to found behavioral economics, and later, with Cass Sunstein, popularized "nudges" as a tool for policy. Kahneman received the Nobel Memorial Prize in Economic Sciences in 2002 and Thaler in 2017.

Methods

  • Observational studies record behavior as it occurs, without intervention: field observation, diary studies, analysis of existing records. They are high in realism and cannot establish cause.
  • Correlational studies measure two or more variables in the same people and test whether they covary. Surveys and individual-difference research are mostly correlational. Correlation does not establish direction or rule out third variables.
  • Experiments manipulate an independent variable and randomize everything else, which is what licenses a causal conclusion. Within-subject designs expose each participant to every condition; between-subject designs assign each participant to one. Most of the paradigms you can run online, from the Stroop test to economic games, are experiments in this sense.

Validity, reliability and replication

A measure is reliable if it gives consistent results: the same person scores similarly on two occasions (test-retest), or the two halves of a task agree with each other (internal consistency). It is valid if it measures what it claims to. Internal validity is the degree to which an effect can be attributed to the manipulation rather than to confounds; external validity is the degree to which it generalizes beyond the sample and setting. Well-known reaction-time effects are often highly reliable at the group level yet surprisingly poor at measuring individual differences, a point made forcefully by Hedge, Powell and Sumner (2018).

The replication crisis of the 2010s, marked by the Open Science Collaboration's 2015 attempt to replicate 100 psychology studies, of which well under half produced significant results in the same direction, changed everyday practice. Preregistration of hypotheses and analysis plans, larger samples, sharing data and code, and reporting effect sizes with confidence intervals are now expected in much of the field. Online data collection contributed by making large samples affordable.

How jsPsych and Cognition fit into the workflow

A typical study today goes through design, preregistration, programming, piloting, data collection, analysis and sharing. jsPsych, the open-source JavaScript library introduced by de Leeuw (2015), covers the programming step: an experiment is a timeline of trials, each handled by a plugin, and the result is a data table with one row per trial. Because the code is plain JavaScript, it can be shared alongside the preregistration and rerun by anyone.

Cognition covers hosting and data collection. You paste jsPsych code (versions 6.0.5 to 8.2.3) into an editor with live preview, upload stimuli, and get a participant link. Data is saved to the server trial by trial and buffered locally if the connection drops; informed consent is configurable; single-use links and recruitment limits control who takes part and how many times; between-subject conditions are balanced on the server and exposed as window.CONDITION; and the results can be exported as CSV or JSON with participant identifiers from Prolific, SONA or MTurk attached to every row. Collaborators can be added per task with viewer, editor or owner roles, which suits shared lab projects and supervised student work.

For concrete illustrations of the field, see ten examples of behavioral science, and for a hands-on start, the Stroop tutorial. You can create a free account to host up to four tasks with sixty participants each.